1. Introduction

This exercise aims to plot interactive population pyramids on a planning area level from year 2000-2010.

Import the packages needed.

packages = c('tidyverse','readxl','dplyr','ggplot2','plotly','gganimate','ggiraph')
for(p in packages){
 if(!require(p, character.only = T)){
  install.packages(p)
 }
 library(p,character.only = T)
 }

2. Data Exploration and preparation

Import both population data The data is imported using read_csv() function of readr package. It can be observed that both datasets have same columns, hence, can be combined.

data0010 <- read_csv("respopagesextod2000to2010/respopagesextod2000to2010.csv")
data1120 <- read_csv("respopagesextod2011to2020/respopagesextod2011to2020.csv")

As the analysis aims to analyse the change of population over the years from 2000 to 2020, the two datasets are combined using rbind().

data <- rbind(data0010, data1120)

Group population by planning area and plot a Pareto chart to inspect the population distribution over planning area. It can be observed from the interactive bar chart that some planning areas are having 0 population.

From the Pareto chart, it can also be observed that more than 50% of the population comes from the top 9 planning areas ordered by total population count. As there are a large number of planning areas, for ease of analysis, the focus will be on these 9 planning areas.

data_grouped_pa <- data %>%
  group_by(PA)%>%
  summarise(
    count_by_pa = sum(Pop, na.rm = TRUE)) 

#order the data and add the cumulative percentage column
data_grouped_pa <- data_grouped_pa[order(data_grouped_pa$count_by_pa, decreasing=TRUE),]

data_grouped_pa$cumulative <- cumsum(data_grouped_pa$count_by_pa)/sum(data_grouped_pa$count_by_pa) * 100

#create the scale on secondary y-axis
scale <- tail(data_grouped_pa$cumulative, n=1)/head(data_grouped_pa$count_by_pa, n=1)


p <- ggplot(data_grouped_pa,
            aes(x = reorder(PA, -count_by_pa),y = count_by_pa,  
                 tooltip=count_by_pa)) +
  geom_bar_interactive(stat = "identity", color="black", fill = 'light blue') + 
  geom_path_interactive(aes(y=cumulative/scale, group=1),colour="red", size=0.9) +
  geom_point_interactive(aes(y=cumulative/scale, group=1,tooltip = cumulative),colour="red") +
  scale_y_continuous(
    # Features of the first y-axis
    name = "Population Count",
    # Add second y-axis and specify its features
    sec.axis = sec_axis(~ .*scale, name="% of Total Running Sum of Population Count")
  ) + 
  theme(axis.text.x = element_text(angle = 90)) + 
  labs(title="Pareto Chart", subtitle="Count of Population by planning area",     
       x="Planning Area", y=expression(count_by_pa(~mu~Ah/~mu~s)))


girafe(ggobj = p,width_svg = 7,
       height_svg = 6)

Filter the data for the top 9 planning areas with the most population amount.

filter_list <- data_grouped_pa$PA[1:9]
print(filter_list)

[1] "Bedok"         "Tampines"      "Jurong West"   "Woodlands"    
[5] "Hougang"       "Yishun"        "Ang Mo Kio"    "Choa Chu Kang"
[9] "Sengkang"

data_filtered <- filter(data, PA %in% filter_list)

data_filtered$PopPercent <- data_filtered$Pop/sum(data_filtered$Pop) * 100

3. Plot the population pyramid of the 9 planning areas

p <- ggplot(data_filtered, aes(x = AG, fill = Sex,
                 y = ifelse(test = Sex == "Males", yes = -Pop, no = Pop),frame = 
                    Time,text = paste0("Population:", Pop,"\n",
                                       "Age group:",AG,"\n",
                                       "Time:", Time)))+
  geom_col(position = "identity") +
  scale_y_continuous(labels = abs, limits = max(data_filtered$Pop) * c(-1,1)) +
  scale_x_discrete(breaks=data_filtered$AG)+
  labs(x = "Age", y = "Population Count") +
  coord_flip() + #theme(legend.position = "none")+
  facet_wrap(~PA) 
  
ggplotly(p,tooltip = "text")%>%
      layout(                       
             title = "Population Pyramids",
             title.yanchor = "top",
             yaxis = list(           
                title = "Age Group") ,
             autosize = FALSE,width = 700, height = 900
      )

4. Challenges

Subplots of population pyramid are squeezed together, making it unclear to read as the axis labels overlapped each other. The main challenge is to find ways to resize the plot. layout() is used for the resizing.

Untitled

1. Introduction

Import the packages needed.

2. Data Exploration and preparation

3. Plot the population pyramid of the 9 planning areas

4. Challenges